Correlations between motor performance and cognitive functions in children born < 1250 g at school age.

Abstract

Very low birth weight born children manifest a higher prevalence of motor and cognitive impairments than term children. Seventy-four prospectively enrolled children born < 1250 g underwent testing of motor (Zurich neuromotor assessment ZNA: timed motor performances and associated movements) and cognitive functions (Kaufman-ABC) at age six years. Children with cerebral palsy or mental retardation were excluded. Adaptive motor tasks (pegboard and dynamic balance) and visuomotor cognitive functions were specifically impaired, and a distinct correlation pattern between motor and cognitive abilities was detected. The adaptive fine motor task (pegboard) correlated with visuomotor functions of the Kaufman-ABC ("triangles", r = 0.35; "matrix analogies", r = 0.39), while pure motor tasks of the ZNA (repetitive, alternating, and sequential movements) did not in spite of impaired motor performance. Timed motor performance below the 10th percentile correlated strongly with cognitive delay (IQ < 85: adaptive fine motor: OR 6.0 [95% CI] 4.7-7.3; adaptive gross motor: OR 7.0 [CI 5.6-8.4]; static balance: OR 9.6 [CI 8.2-11.0]). In conclusion, motor deficits in children born < 1250 g without severe disabilities correlate with specific cognitive impairments, in particular of the visuomotor domain. The correlation pattern may indicate specific dysfunction in visuomotor transformation, the intermediate process between visual-perceptual input and motor output. Early assessment of both motor and cognitive functions using standardized assessment tools is important to determine the extent and combination of specific developmental disturbances and to tailor therapeutic intervention.

Full text

Page 1

Abstract
Very low birth weight born children manifest a higher preva-
lence of motor and cognitive impairments than term children.
Seventy-four prospectively enrolled children born <1250g
underwent testing of motor (Zurich neuromotor assessment
ZNA: timed motor performances and associated movements)
and cognitive functions (Kaufman-ABC) at age six years. Children
with cerebral palsy or mental retardation were excluded. Adap-
tive motor tasks (pegboard and dynamic balance) and visuomo-
tor cognitive functions were specifically impaired, and a distinct
correlation pattern between motor and cognitive abilities was
detected. The adaptive fine motor task (pegboard) correlated
with visuomotor functions of the Kaufman-ABC (“triangles”,
r=0.35; “matrix analogies”, r=0.39), while pure motor tasks of
the ZNA (repetitive, alternating, and sequential movements) did
not in spite of impaired motor performance. Timed motor per-
formance below the 10th percentile correlated strongly with
cognitive delay (IQ <85: adaptive fine motor: OR 6.0 [95% CI]
4.7–7.3; adaptive gross motor: OR 7.0 [CI 5.6–8.4]; static bal-
ance: OR 9.6 [CI 8.2–11.0]). In conclusion, motor deficits in chil-
dren born <1250g without severe disabilities correlate with
specific cognitive impairments, in particular of the visuomotor
domain. The correlation pattern may indicate specific dysfunc-
tion in visuomotor transformation, the intermediate process be-
tween visual-perceptual input and motor output. Early assess-
ment of both motor and cognitive functions using standardized
assessment tools is important to determine the extent and com-
bination of specific developmental disturbances and to tailor
therapeutic intervention.
Key words
Motor performance · visuomotor · cognition - correlation - very
low birth weight
Abbreviations
Kaufman-ABC Kaufman assessment battery for children
MPCmental processingcompositeof the Kaufman-ABC
ORodds ratio
SESsocioeconomic status
VLBWvery low birth weight
Zurich NMAZurich neuromotor assessment
Introduction
Because the survival rates of very low birth weight (VLBW) chil-
dren have substantially improved over the past decades [18,27],
the focus of research has shifted towards determining long-term
outcomes and impairments. The prevalence of major disabilities
such as cerebral palsy or severe mental retardation has been re-
ported at a relatively constant level between 7 and 20% [5,11].
Moderate and minor motor disabilities occur in 30 to 40% of
VLBW children [12,26]. Childrenmayalso suffer fromminor cog-
nitive impairments, often not diagnosed before school age [2].
These deficits include learning disabilities, borderline intellec-
tual functioning, behavioural problems, attention-deficit/hyper-
activity disorders, and specific neuropsychological impairments
such as visuospatial and visuomotor impairments [2,6]. These
Original Article
6
Affiliation
Child Development Centre, University Children’s Hospital Zurich, Switzerland
Correspondence
B. Latal Hajnal · Child Development Centre · University Children’s Hospital Zurich · Steinwiesstraße 75 ·
8032 Zurich · Switzerland · E-mail: bea.latal@kispi.unizh.ch
Received: April 12, 2005 · Accepted after Revision: December 23, 2005
Bibliography
Neuropediatrics 2006; 37: 6–12 © Georg Thieme Verlag KG Stuttgart · New York ·
DOI 10.1055/s-2006-923840 ·
ISSN 0174-304X
J. Seitz
O. G. Jenni
L. Molinari
J. Caflisch
R. H. Largo
B. Latal Hajnal
Correlations between Motor Performance
and Cognitive Functions
in Children Born <1250g at School Age

Page 2

so-called “high prevalence/low-severity dysfunctions” are esti-
mated to occur in almost 50% of VLBW children [9,20].
Several authors have raised the question whether deficits in mo-
tor and cognitive abilities are related [6,20]. In fact, visuomotor
skills are often impaired in VLBW children and have been asso-
ciated with deficits in fine motor performance [9]. A more de-
tailed evaluation of the motor-cognition relationship and visuo-
motor functions may have implications regarding developmen-
tal assessment and therapeutic strategies.
In this study, we examined motor and cognitive functions and
specific correlation patterns in six-year-old children born
<1250g withoutcerebral palsy ormental retardation. TheZurich
neuromotorassessment, a standardised motor test, was correlat-
ed with subtests of the Kaufman assessment battery for children
(Kaufman-ABC) assessing cognitive function.
Materials and Methods
Population
The population was prospectively recruited from the Maternity
Hospital Zurich, a tertiary care centre serving a population of ap-
proximately 1.7 million people. Between July 1992 and June
1994, all live born infants with a birth weight below 1250 gram
born at the Maternity Hospital were enrolled. The study was ap-
proved by the local ethics committee. Of 146 live-born infants,
45 (31%) died during the neonatal period. No child died after dis-
charge from the nursery. None of the surviving infants had chro-
mosomal disorders or severe malformations.
Neurodevelopmental assessment was performed at three, nine,
24 months and at three and six years corrected age. Of the 101
surviving infants, 88 (87%) returned for follow-up at six years of
age. Thirteenchildrenwere lost to follow-up because the parents
refusedtoreturnorcould not be traced. Onechild was brought to
the six-year examination but could not be examined due to be-
havioural abnormalities. Perinatal variables did not differ be-
tween the children with and without follow-up examination.
The socioeconomic status (SES) [11] of those parents who re-
turned was higher than of those who did not (Mann-Whitney,
p=0.001). Five of the lost children were examined at age two
years; they had normal mental developmental indices on the
Bayley scales of infant development II and none had cerebral
palsy.
Children with cerebral palsy [12] (n=7) or with mental retarda-
tion (IQ<70, n=6) were excluded from the analysis; thus motor
and cognitive outcomes were analysed for children who were
able to complete cognitive and motor testing with the same test
batteries (n=74).
Children were examined at the Child Development Centre of the
University Children’s Hospital Zurich at a mean age of six years
(range 5.4–7.4 years). Mean SES (estimated on a 12-point scale
based on paternal occupation and maternal education [11]);
score of 2 = highest, score of 12 = lowest socioeconomic level)
was 5.7 (range 2–11). Perinatal characteristics of the 74 children
are summarised inTable 1. Bronchopulmonary dysplasia was de-
fined as oxygen supplementation at day 28 of life. Intraventricu-
lar haemorrhage was graded according to Papile [25] and leuko-
malacia according to de Vries [4]. Cystic periventricular leuko-
malacia was diagnosed in two children and was grade 2. They
both had minor neurological abnormalities, and their IQs were
102 and 98, respectively. One child had an intraventricular
haemorrhage grade 3. He had mild neurological abnormalities
and his IQ was 82.
Zurich neuromotor assessment (Zurich NMA)
The Zurich NMA is a standardised and reliable testing procedure
inwhichdistinct motortasks areassessed fortimedperformance
and movement quality. Age- and gender-specific normative data
are based on 662 normally developed children and adolescents
between ages of 5 and 18 years [14–16]. Children were recruited
from the local area of Zurich.
The testing procedure includes the assessment of pure motor
tasks, adaptivefine motor tasks, adaptivegrossmotor tasks, stat-
ic balance, and stress gaits. Some motor tasks (sequential finger
movements, static and dynamic balance) are performed with in-
creasing difficulty and depend on the child’s age. The following
descriptions are given for the age of six years. The pure motor
tasks consist of repetitive foot (20 per foot), hand (20 per hand),
and finger movements (20 per hand), alternating foot (10 per
foot) and hand movements (10 per hand), and sequential move-
ments of the fingers (three sequences of finger-thumb opposi-
Table 1 Perinatal variables and outcome for children born <1250g
Perinatal variables
Birth weight (gram)
M (SD)1010 (145)
Gestational age (weeks)
M (range)28.6 (25.7–33.1)
Small for gestational age (< P 10)
N (%) 28 (37.3)
Sex (male/female)
N32/42
Twins/triplet
N (% total)20/7 (36.0)
Maternal age (years)
M (range)30.4 (20.0–42.2)
5-minute Apgar
M (SD)7.6 (1.5)
Arterial cord blood pH
M (SD)7.27 (0.07)
Sepsis (culture positive)
N (%) 18 (24)
Retinopathy of prematurity
≥ grade 3
Bronchopulmonary dysplasia
N (%) 4 (5.3)
N (%)34 (45.3)
Patent ductus arteriosus
N (%) 4 (5.3)
Necrotising enterocolitis
N (%)6 (8.0)
Intraventricular haemorrhage 1–4
N (%)31 (41.3)
Intraventricular haemorrhage 3
N (%)1 (1.3)
Cystic periventricular leukomalacia
N (%)2 (2.7)
Echodensities >7 days
N (%)16 (21.3)
Neurosensory impairment at six years
Major visual impairment
(acuity <0.3)
N (%)1 (1.3)
Glasses (for strabism/refractive
error)
N (%)10 (13.0)
Bilateral hearing aids
N (%)2 (2.7)
M = mean, N = number, SD = standard deviation
Seitz J et al. Correlations between Motor … Neuropediatrics 2006; 37: 6–12
Original Article
7

Page 3

tion per hand). Adaptive tasks require the integration of other
sensory systems (i.e., visual or tactile-kinaesthetic): in the adap-
tive fine motor task “pegboard”, 12 plastic pegs are placed in
holes,first with thedominant thenwiththe non-dominanthand.
The passive hand is placed in a relaxed position beside the peg-
board. The “pegboard” is performed twice. The adaptive gross
motor task “dynamic balance” consists of five double jumps side-
ways and of two lengths forward jumps within two lines. The
“static balance” task involves standing on one foot.
The child’s performance is videotaped. Motor performance is
timed with a stopwatch. Movement quality is scored from video-
tape recordingsbyclassifying frequency (score0–10)and degree
(score 0–3) of associated movements during timed performance
and stress gaits of the contralateral and ipsilateral extremity, the
face and head/body. A detailed description of the Zurich NMA is
given in manual of the Zurich NMA [16].
The results of the Zurich NMA are expressed as components [7].
The pure motor component combines the timed performance of
all repetitive, alternating, and sequential tasks from both ex-
tremities. The adaptive fine motor component combines all
timed performances of the pegboard. The adaptive gross motor
component combines the results of the motor tasks of both dy-
namic balances, and the static balance component those of the
static balance. The associated movements component expresses
the sum ofall associatedmovementsof all motortasks and stress
gaits. Intra-rater reliability for timed performance of the compo-
nents pure motor, adaptive fine motor, gross motor, and static
balance is 0.94, 0.99,1.00,and1.00,respectively, andfortheasso-
ciated movements component 0.82. Inter-rater reliability for the
same components is 0.89, 0.97, 0.89, 0.98, and 0.78, respectively.
Children were examined during a full morning or afternoon. The
orderof the examinationwas: joining phase, personal and family
history, cognitive testing (Kaufman-ABC), break, motor testing
(Zurich NMA), general physical and neurological examination.
Three experienced developmental paediatricians not involved in
the neonatal care of the study subjects performed all examina-
tions. The children were in good physical health and had no
chronic pulmonary or other disorders. Examiners were regularly
supervised by video recordings. One trained research assistant
(J.S.) rated associated movements from video recordings. Inter-
rater reliability was assessed for 20 childrenwith an experienced
instructor who was involved in the development of the Zurich
NMA (J.C.). Kappa coefficient was 0.82.
Kaufman-ABC
Cognitive functions were assessed using the German version of
the Kaufman-ABC [22]. The Kaufman-ABC is an intelligence test
based on information-processing theory of cognitive develop-
ment [13]. It is composed of two parts, the mental processing
composite (MPC) and the achievement scale.
The MPC comprises tasks reflecting experience-independent
problem-solving strategies and intellectual capabilities of the
child. The MPC is further subdivided into the sequential process-
ing scale consisting of the subtests “hand movements”, “number
recall” and “word order”, and the simultaneous processing scale,
consisting of the subtests “magic window”, “face recognition”,
“gestalt recognition”, “triangles”, “matrix analogies”, “spatial
memory”, and “photo series”. Both scales reflect different brain-
inherent modes of cognition (for details see [13]).
The achievement scale contains learning-dependent tasks and
measures acquired knowledge and application of skills. The
achievement scale consists of “faces and places”, “arithmetic”,
“riddles”, and “reading/decoding”. Certain subtests (“magic win-
dow”, “face recognition”, “reading/decoding”, and “reading/
understanding”)areadministeredwithin definedagerangesthat
were below or above the age of six years and thus were not test-
ed. In the German version of the Kaufman-ABC [21], scoring of
the subtest “faces and places” is not reliable for Swiss children
because of cultural differences between Germany and Switzer-
land. Thus, this subtest had to be deleted from analysis.
Data analysis
Performance in the Zurich NMA is expressed as standard devia-
tion from the mean of the reference population (Z score) accord-
ing to age and sex. A negative Z score reflects performancebelow
average (slow timed performance, frequent and strong associat-
ed movements), a positive Z score performance above average
(fast timed performance or few and weak associated move-
ments).
Somechildrenwerenotable tocompleteall motortasks and cog-
nitive subtests due to clumsiness (motor test) or fatigue. Thus,
total numbers may vary (Tables 1, 2, and 3). Because the results
Table 2 Motor performance (Z scores) of children born <1250g
Zurich NMAN*Median (range)N (%) < P10p
Timed performances
Pure motor tasks
72–0.45 (–2.72/1.58) 18 (24%)<0.01
Adaptive fine motor
74–0.73 (–5.24/2.04) 28 (38%) <0.001
Adaptive gross motor
72–0.82 (–4.14/2.74)19 (26%)<0.001
Static Balance
70–0.45 (–3.53/2.07) 13 (18%)<0.01
Associated movements
68–1.90 (–3.95/–0.40)50 (68%) <0.001
Zurich NMA = Zurich neuromotor assessment, N (%) < P10 = number (percent) below 10th percentile, p value (Wilcoxon sign test): difference from reference population.
*Some children could not perform all motor tasks of the Zurich NMA.
Seitz J et al. Correlations between Motor … Neuropediatrics 2006; 37: 6–12
Original Article
8

Page 4

of the Zurich NMA were not normally distributed in the studied
population, differences in motor performance between children
born <1250g and the reference population were reported using
non-parametric tests. Results from the Kaufman-ABC were nor-
mally distributed and a two-tailed standard t-test was used to
determine differences from the norm. Spearman rank correla-
tions between motor and cognitive performance were calculat-
ed. When applying the Bonferroni correction factor of 50, corre-
lations with a p<0.001 before correction remained significant.
This estimate, however, was too conservative. To provide full in-
formation, all correlations are shown with uncorrected p values.
Two to three correlations with p<0.05 could occur by chance
when calculating 50 correlations. We found 12 significant corre-
lations which were considered to occur beyond chance. The null
hypothesis that all 50 correlations would be 0 could be rejected
at the p<0.05 level [23].
Sepsis, male gender, multiple birth status, small for gestational
age status, and periventricular leukomalacia were not associated
with motor or with cognitive outcome. Perinatal and demo-
graphic variables associated with outcome (intraventricular
haemorrhage with MPC [p=0.01, univariate analysis] and with
dynamic balance component [p=0.02], patent ductus arteriosus
with MPC [p=0.04], SES with MPC [p=0.001], and bronchopul-
monary dysplasia with dynamic balance component [p=0.07])
were included in the multiple regression analysis to determine
the influence of risk factors on the correlation between motor
and cognitive outcome. We calculated odds ratios to estimate
the risk of poor cognitive function (MPC <85, <7 for subtests) in
association with impaired motor performance (<10th percen-
tile).
Results
Outcome at age six years
One child had a major visual impairment, but could completethe
Zurich NMA, ten children required corrective glasses either for
strabism or refractive errors (Table 1). Three children were diag-
nosed with language developmental disorder and required spe-
cial schooling. Minor language problems (articulation) or mild
language delay occurred in 24 (32%) children.
Children with visual problems (glasses or major visual impair-
ment) performed more poorly on the pure motor and the adap-
tive fine motor component (Mann-Whitney p=0.03 and 0.02, re-
spectively).No significant differencewas found for the Kaufman-
ABC test (overall and subtest level).
Motor performance
Neuromotor performanceis presented inTable 2. Performancein
all time components was significantly lower compared to the
reference population. Among the four time components, the
adaptive fine (pegboard) and gross motor (dynamic balance)
components were particularly affected. The associated move-
ments component was the most affected component.
Cognitive function
Mean MPC of the Kaufman-ABC was 95.5 (Table 3). Performance
in the sequential processing scale was poorer than in the simul-
taneous processingscale (p=0.007, t-testfordependentsample).
A distinct pattern of impairment was found on the subtest level.
In thesequential processingscale,“handmovements”and“num-
ber recall” were below the normal range (0.001 and <0.001, re-
spectively). In the simultaneous processing scale, mean scores
in “triangles”, a visuomotor item [13], were significantly
(p<0.001) belowthe norm. Performancein “gestalt recognition”,
“spatial memory”, and “photo series”, visual perceptual items
Table 3 Cognitive performance of children born <1250g
Kaufman-ABC N*MeanSD RangeN (%) < –1 SDp value+
Mental processing composite
7495.510.370–125 10 (14%)<0.001
Sequential processing scale
7493.212.272–12119 (26%)<0.001
Hand movements
748.8 2.8 4–1616 (21%) 0.001
Number recall
747.9 2.33–1427 (36%)<0.001
Word order
739.62.65–165 (7%) n.s.
Simultaneous processing scale
7498.813.3 64–136 12 (16%)n.s.
Gestalt recognition
73 10.42.64–168 (11%) n.s.
Triangles
748.7 2.52–1312 (16%)<0.001
Matrix analogies
74 9.62.84–194 (5%)n.s.
Spatial memory
73 9.92.63–14 9 (12%)n.s.
Photo series
62 9.93.35–19 14 (22%)n.s.
Achievement scale
7290.117.1 56–122 12 (16%)<0.001
Arithmetic
7298.7 16.764–145 15 (20%) n.s
Riddles
73 92.318.351–120 18 (22%)0.001
Kaufman-ABC=Kaufmanassessmentbattery forchildren,N (%)<1–SD =number(percent)below–1standard deviation (<85or <7),+univariate t-test comparedtonorm
(n.s.=not significant). *Some children could not complete all subtests of the Kaufman-ABC.
Seitz J et al. Correlations between Motor … Neuropediatrics 2006; 37: 6–12
Original Article
9

Page 5

[14], was best. The mean scorein thesubtest “riddles”was signif-
icantly below normal. Children performed within the normal
range on the subtest “arithmetic”.
Relationship between motor performance and
cognitive function
Three time components significantly correlated with the MPC
(Table 4): the adaptive fine motor component, the adaptive gross
motor, and the static balance component, whereas the pure mo-
tor and associated movements component did not. On the sub-
test level, the strongest correlations were found between the
adaptive fine motor component and “triangles”, “matrix analo-
gies” and “spatial memory”, between the adaptive gross motor
component and “number recall” as well as “triangles”. The static
balance componentcorrelatedwithspatial memoryand thepure
motor component with “hand movement” and “spatial memo-
ry”. The associated movements component did not correlate
with the MPC (r=0.12, p=0.32) and on a subtest level only with
“spatial memory”.
The Kaufman-ABC subtest “spatial memory” significantly cor-
related with all time components as well as with the associated
movements component. All correlations retained their level of
significance when perinatal variables were controlled for, except
for the correlation between the adaptive gross motor component
and thesimultaneous processingscale and within that scalewith
the subtest “photo series”. We calculated odds ratios (OR) to ex-
press the risk of poor cognitive function (MPC <85) in associa-
tion with impaired motor performance (<10th percentile). The
adaptive fine motor (OR 6.0, 95% confidence interval [CI] 4.7–
7.3), the adaptive gross motor (OR 7.0, CI 5.6–8.4), and the static
balance component(OR 9.6, CI 8.2–11.0) were significantly asso-
ciated with an MPC below 85, whereas the pure motor compo-
nent was not (OR 1.9, CI 0.5–3.4).
Discussion
In this study, adaptive motor tasks (pegboard and dynamic bal-
ance) assessed by the ZNA and visuomotor cognitive functions
examined in subtests of the Kaufman-ABC were specifically im-
paired, and a distinct correlation pattern between motor and
cognitive abilities was detected in children born <1250g with-
out major handicaps.
The findings of our study provide new insights into existing
knowledge of visuomotor impairment in VLBW or very preterm
children [6,9]. The impairment and correlation pattern between
motor and cognitive function may indicate which part of the vi-
suomotor system is affected, that is visual-perceptual input, mo-
tor output, or intermediate processes between input and output
such as visuomotor transformation. We found that adaptive fine
motor and partly gross motor tasks correlated with visuomotor
subtests of the Kaufman-ABC, whereas the pure motor compo-
nent of the ZNA did not despite impaired performance. Thus, we
suggest that visuomotor impairments in preterm children with-
out cerebral palsy or mental retardation are probably not deter-
mined by motor deficiency (motor output).
The visuomotor impairment was confirmed in other examined
domains. Among the four timed components of the Zurich
NMA, adaptive tasks (pegboard and dynamic balance) were par-
ticularly impaired. They require the integration of other sensory
systems (i.e., visual or tactile-kinaesthetic) and reflect visuomo-
Table 4 Correlations between Zurich NMA and Kaufman-ABC
Zurich NMA
component
Pure motorAdaptive
fine motor
Adaptive
gross motor
Static balanceAssociated
movements
n=72n=74 n=72 n=70 n=68
Kaufman-ABC
Mental processing composite
0.220.41***
0.37**
0.36**
0.12
Sequential processing scale
0.210.190.30*
0.22 0.10
– hand movements
0.35**
0.11 0.120.20 0.23
– number recall
0.090.150.41***
0.06–0.04
– word order
0.10 0.26*
0.27*
0.22 0.07
Simultaneous processing scale
0.15 0.40***
0.30*
0.35**
0.03
– gestalt recognition
0.03 0.200.170.05–0.03
– triangles
0.19 0.35**
0.30*
0.210.10
– matrix analogies
–0.030.39***
–0.060.20–0.07
– spatial memory
0.33**
0.38***
0.29*
0.38**
0.28*
– photo series
0.020.220.27*
0.32*
–0.05
Achievement Scale
–0.06 0.14 0.180.17–0.15
– arithmetic
0.050.180.12 0.25*
–0.08
– riddles
–0.130.100.24*
0.04–0.15
Spearman rank correlation coefficients *p<0.05, **p<0.01, ***p<0.001. Kaufman-ABC = Kaufman assessment battery for children, Zurich NMA = Zurich neuromotor
assessment
Seitz J et al. Correlations between Motor … Neuropediatrics 2006; 37: 6–12
Original Article
10